International Journal of Trend in Scientific Research and Development (IJTSRD)

Special Issue: International Conference on Advances in Engineering, Science and Technology – 2021
Organized by: Uttaranchal Institute of Technology, Uttaranchal University, Dehradun
Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470

Covid-19 Prediction in India using Machine Learning

Sarfraj Alam1, Vipul Kumar1, Sweta Singh1, Sweta Joshi1, Madhu Kirola2
1Student, 2Assistant
Professor,
1,2Computer Science and Engineering, Uttaranchal Institute of Technology, Dehradun, Uttarakhand, India

ABSTRACT How to cite this paper: Sarfraj Alam |

Various computational models are used around the world to predict the Vipul Kumar | Sweta Singh | Sweta Joshi |
number of infected individuals and the death rate of the COVID-19 outbreak Madhu Kirola "Covid-19 Prediction in
[3]. Machine learning-based models are important to take proper actions. Due India using Machine Learning" Published
to the ample of uncertainty and crucial data, the aerodynamic models have in International Journal of Trend in
been challenged regarding higher accuracy for long-term prediction of this Scientific Research and Development
disease [1]. By researching the COVID19 problem, it is observed that lockdown (ijtsrd), ISSN: 2456-
and isolation are important techniques for preventing the spread of COVID-19 6470, Special Issue |
[2]. In India, public health and the economical condition are impacted by International
COVID-19, our goal is to visualize the spread of this disease [5]. Machine Conference on
Learning Algorithms are used in various applications for detecting adverse Advances in
risk factors. Three ML algorithms we are using that is Logistic Regression (LR), Engineering, Science
Support Vector Machine (SVM), and Random Forest Classifier (RFC). These and Technology – IJTSRD42458
machine learning models are predicting the total number of recovered 2021, May 2021,
patients as per the date of each state in India [8]. pp.8-12, URL:
www.ijtsrd.com/papers/ijtsrd42458.pdf
KEYWORDS: Data Analysis and Visualization, Logistic Regression, Support Vector
Machine, Random Forest Classifier Copyright © 2021 by author(s) and
International Journal of Trend in Scientific
Research and Development Journal. This
is an Open Access article distributed
under the terms of
the Creative
Commons Attribution
License (CC BY 4.0)
(http://creativecommons.org/licenses/by/4.0)

INTRODUCTION:
Various drastic respiratory syndrome coronavirus 2, also economical condition of the people. The research on the
called SARS-CoV-2, is reported as a strain of virus inducing household and international endemic and the upcoming
the respiratory disease of COVID-19 The coronavirus disease evolution shift has been subject undergoing intense study
is confirmed to be extremely contagious by the global [6]. While no model can precisely foresee the rates of
nations and the World Health Organization (WHO). The infection and mortality, there have been efforts made to
COVID-19 pandemic has become an international concern contemplate and inspect the strengths and drawbacks of
that has recognized as a public health emergency all over the severe studies and models conferred regarding the
world. Epidemiological models have been broadly used by coronavirus.
media and officials to assess the outbreak, recognize the
Background:
peak beforehand, and also anticipate the mortality rate.
We investigate the matter of modeling the latest coronavirus
Outburst prediction models have manifested to be elemental
(Covid19) spread the constrained or practical scenarios in
to furnish observations into the harm brought by COVID-19
India. It is monitored that lock-down and isolation are the
[3]. Moreover, the prediction models are taken as a scientific
crucial steps to control the spread of this disease. We study
report to construct new plans and to assess the conditions of
the result of those prevention techniques mathematically
curfew. The COVID-19 pandemic has been broadcasted to be
and build a model to predict the new cases or total infected
exceedingly aggressive to spread [1]. Towards the end of
cases. This prediction is required to organize reports and to
2019, the new coronavirus spread widely in China, and a
take further steps to stop preventing this disease. [2].
huge mass get infected. At present, the household upsurge
COVID-19 has been proved a global pandemic and a couple
has been productively managed, although the new
of tests are using various ML based models to manage the
coronavirus is escalating speedily in other areas. Presently,
reports of the various cases at global level. Here, we
Europe has become the base of the ongoing outburst of new
showcase our model that would be able to predict the arrival
pneumonia. On March 11, the World Health Organization
of new cases across India. Various classification methods are
declared new pneumonia outburst a "global pandemic" [5].
used for making an crucial report on COVID-19 across India
This disease has provoked huge impact to the health and
[4].

@ IJTSRD | Unique Paper ID – IJTSRD42458 | ICAEST-21 | May 2021 Page 8

 Special Issue: International Conference on Advances in Engineering, Science and Technology – 2021 (ICAEST-21)
Available online @ www.ijtsrd.com eISSN: 2456-6470
Machine learning-based models have showcase the infected cases, death rates, and the recovered cases
importance to predict the perioperative output to enhance estimates in the next 10 days across India [8].
the decision making on the longer-term period. Various
Proposed Methodology:
prediction techniques are used to handle speculate
We extracted Indian country data about the COVID-19
problems. The study of COVID-19 shows the potential of ML
disease from the Kaggle website Data was extracted on 4
based models to predict the number of upcoming patients
September 2020. Using Jupyter Notebook and Python
suffering from COVID-19 [1].
libraries we merged and filtered the final analysis of the
Computational techniques for machine learning-based dataset. In this, we use Pandas, NumPy, matplpotlib, seaborn
models have shown their importance for monitoring the and, sklearn library. From the list of unique states, we get the
potential threats for valueable decisions. Machine learning names of the top five states having maximum confirmed
algorithms utilized in various applications show the cases, totally cured cases, and total new cases. Then we plot
inauspicious risk factors. The study shows the ability of the facet grid for total confirmed cases concerning the state.
model to estimate the number of individuals who are We plot one feature map which helps us to do the feature
affected by this disease. This machine learning based model selection. And based on feature selection, we divide the
has different kinds of outcomes like the number of newly dataset into training and test data.

@ IJTSRD | Unique Paper ID – IJTSRD42458 | ICAEST-21 | May 2021 Page 9

 Special Issue: International Conference on Advances in Engineering, Science and Technology – 2021 (ICAEST-21)
Available online @ www.ijtsrd.com eISSN: 2456-6470
Here are some visualizations for total confirmed cases, cured cases, new cases, and facetgrid for each state (Confirmed Case) –

Confirmed Cases

Cured Cases

@ IJTSRD | Unique Paper ID – IJTSRD42458 | ICAEST-21 | May 2021 Page 10

 Special Issue: International Conference on Advances in Engineering, Science and Technology – 2021 (ICAEST-21)
Available online @ www.ijtsrd.com eISSN: 2456-6470

New Cases

Facetgrid of Each State (Confirmed Cases)

Feature Map

@ IJTSRD | Unique Paper ID – IJTSRD42458 | ICAEST-21 | May 2021 Page 11

 Special Issue: International Conference on Advances in Engineering, Science and Technology – 2021 (ICAEST-21)
Available online @ www.ijtsrd.com eISSN: 2456-6470
We use three machine learning algorithms to build the [4] R. Sujath1 • Jyotir Moy Chatterjee2 • Aboul Ella
model: Hassanien3 - A machine learning forecasting model
for COVID-19 pandemic in India - SpringerVerlag
Logistic Regression: It is an classification based algorithm
GmbH Germany - 30 May 2020.
which is used to predict the unique values. It predict the
output in the form of binary variable Basically it predicts [5] Gina Cahill MPH a, Carleigh Kutac MPH b, Nicholas L.
whether the particular event is going to happen or not. Its Rider DO - Visualizing and assessing US county-level
output falls in then range between 0 and 1 [6]. COVID19 vulnerability - American Journal of Infection
Control.
Support Vector Machine: It is an classification based [6] Bernhard O. Josephusa, Ardianto H. Nawirb, Evelyn
method in which we plot item as an point in N-dimensional Wijayac, Jurike V. Moniagad *, Margaretha Ohyvere -
world (N denotes total number of features the dataset Predict Mortality in Patients Infected with COVID-19
contains). Each point has two co-ordinates that is X and Y Virus Based on Observed Characteristics of the
and this co-ordinates are called as support vectors [8]. Patient using Logistic Regression – Procedia.
Random Forest Classifier: The random forest classifier is a [7] Mainak Adhikari a,∗, Ambigavathi Munusamy b -
powerful algorithm for disease prediction and it is a iCovidCare: Intelligent health monitoring framework
collection of decision trees. It takes the votes from all the for COVID-19 using ensemble random forest in edge
decision trees depending on the parameters we give and it networks - Internet of Things - Accepted 24 February
combines all the votes of decision trees to produce a single 2021.
output. These decision trees contains two different stages [8] Ramesh Kumar Mojjada a,⇑, Arvind Yadav a, A.V.
that is ensemble learning and feature engineering. First, it Prabhu b, Yuvaraj Natarajan c - Machine learning
maps from the training data that is uniquely selected, which models for covid-19 future forecasting - Materials
ensures that each tree predicts some values using machine Today: Proceedings - Accepted 24 October 2020.
learing protocols. After this, it collects all the predicted [9] (https://www.kaggle.com/imdevsk p/covid19-
values to produce the best unique output on the basis of corona-virus-indiadataset?select=complete.csv)
votes [7].
[10] https://www.geeksforgeeks.org/python-pandas-
Result Analysis: In this study, we build a computational timestamp-toordinal/
model that will predict the total number of recovered [11] Kumar, R., Singh, G., Joshi, K. Emotion Recognition
patients that are suffering from COVID-19. The dataset we System using Local Binary Pattern. International
used contains the record about total confirmed cases, total Journal of Inventive Engineering and Sciences (IJIES)-
deaths, cured patients, new cases, and total recovered Volume, 5.
patients per state wise. Three machine learning algorithms [12] Joshi, K., Joshi, K, N., Diwakar, M., Tripathi, N, A.,
are used to predict the number of recovered patients. Gupta, H. Multi-Focus Image Fusion using Non-Local
S.NO Algorithm Accuracy Mean Filtering and Stationary Wavelet Transform.
1. Logistic Regression 36.74% International Journal of Innovative Technology and
2. Support Vector Machine 36.74% Exploring Engineering (IJITEE)-Volume,9, 344-350.
3. Random Forest Classifier 36.10% [13] Joshi, Kapil and Kirola, Madhu and Chaudhary, Sumit
and Diwakar, Manoj and Joshi, N.K., Multi-Focus
Conclusion: We have looked into the issue of Covid19 Image Fusion Using Discrete Wavelet Transform
spread in India in a big picture. A mathematical model has Method (March 14, 2019). International Conference
been accepted, which accompanies the actual data course of on Advances in Engineering Science Management &
Covid19 spread in India. As we do lots of data visualization Technology (ICAESMT) - 2019, Uttaranchal
to see in the particular region how many cases are arising University, Dehradun, India. Available at SSRN:
day by day and after that making use of the machine learning https://ssrn.com/abstract=3383141 or
algorithms we will train our model in such a way that it will http://dx.doi.org/10.2139/ssrn.3383141
predict up to maximum accuracy it can do.
[14] Gupta, Himanshu and Chauhan, Himanshu and
References: Bijalwan, Anchit and Joshi, Kapil, A Review on Image
[1] FURQAN RUSTAM 1, AIJAZ AHMAD RESHI 2, Denoising (March 14, 2019). International Conference
(Member, IEEE), ARIF MEHMOOD 3, SALEEM ULLAH on Advances in Engineering Science Management &
1, BYUNGWON ON4, WAQAR ASLAM 3, (Member, Technology (ICAESMT) - 2019, Uttaranchal
IEEE), AND GYU SANG CHOI - COVID-19 Future University, Dehradun, India. Available at SSRN:
Forecasting Using Supervised Machine Learning https://ssrn.com/abstract=3385374 or
Models – IEEE Access - Accepted May 13, 2020. http://dx.doi.org/10.2139/ssrn.3385374
[2] Arti M. K., Senior Member, IEEE and Kushagra [15] Gupta, Asheesh and Ahluwalia, Harshit and Adhikari,
Bhatnagar - Modeling and Predictions for COVID 19 Preeti and Chaudhary, Sumit and Joshi, Kapil, A New
Spread in India – ResearchGate – April 2, 2020. Approach To Transfer Data Using RedTacton
[3] Sina F. Ardabili 1, Amir Mosavi 2,3,*, Pedram Ghamisi Technology (March 15, 2019). International
4, Filip Ferdinand 2, Annamaria R. Varkonyi-Koczy 2, Conference on Advances in Engineering Science
Uwe Reuter 3, Timon Rabczuk 5, Peter M. Atkinson 6 - Management & Technology (ICAESMT) - 2019,
COVID-19 Outbreak Prediction with Machine Uttaranchal University, Dehradun, India. Available at
Learning. SSRN: https://ssrn.com/abstract=3395692 or
http://dx.doi.org/10.2139/ssrn.3395692

@ IJTSRD | Unique Paper ID – IJTSRD42458 | ICAEST-21 | May 2021 Page 12